Table of Contents

1  Ground predictions
1.1  PVLib Clearsky
2  Train/test on NSRDB data to find optimal parameters
3  Train on all NSRDB data, test various freq of ground data
3.1  1 min freq ground data
4  Save model
5  Conclusion

In [1]:
import pandas as pd
import numpy as np
import os
import datetime
import matplotlib
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import make_axes_locatable
from sklearn import tree
from sklearn import ensemble

import pytz
import itertools
import visualize
import utils
import pydotplus
import xgboost as xgb

from sklearn import metrics
from sklearn import model_selection

import pvlib
import cs_detection

import visualize_plotly as visualize
from plotly.offline import download_plotlyjs, init_notebook_mode, plot, iplot
import plotly.graph_objs as go
init_notebook_mode(connected=True)

from IPython.display import Image

%load_ext autoreload
%autoreload 2

np.set_printoptions(precision=4)
%matplotlib notebook

Ground predictions

PVLib Clearsky

Only making ground predictions using PVLib clearsky model and statistical model. NSRDB model won't be available to ground measurements.

In [2]:
nsrdb = cs_detection.ClearskyDetection.read_pickle('abq_nsrdb_1.pkl.gz')
nsrdb.df.index = nsrdb.df.index.tz_convert('MST')
nsrdb.trim_dates('01-01-2014', '01-01-2016')
In [3]:
nsrdb.time_from_solar_noon('Clearsky GHI pvlib', 'tfn')
nsrdb.time_from_solar_noon_ratio2('Clearsky GHI pvlib')
In [4]:
feature_cols = [
'ghi_status',
'tfn',
'abs_ideal_ratio_diff',
'abs_ideal_ratio_diff mean',
'abs_ideal_ratio_diff std',
# 'abs_ideal_ratio_diff max',
# 'abs_ideal_ratio_diff min',
'GHI Clearsky GHI pvlib gradient ratio', 
'GHI Clearsky GHI pvlib gradient ratio mean', 
'GHI Clearsky GHI pvlib gradient ratio std', 
# 'GHI Clearsky GHI pvlib gradient ratio min', 
# 'GHI Clearsky GHI pvlib gradient ratio max', 
'GHI Clearsky GHI pvlib gradient second ratio', 
'GHI Clearsky GHI pvlib gradient second ratio mean', 
'GHI Clearsky GHI pvlib gradient second ratio std', 
# 'GHI Clearsky GHI pvlib gradient second ratio min', 
# 'GHI Clearsky GHI pvlib gradient second ratio max', 
'GHI Clearsky GHI pvlib line length ratio',
'GHI Clearsky GHI pvlib line length ratio gradient',
'GHI Clearsky GHI pvlib line length ratio gradient second'
]

target_cols = ['sky_status']

Train/test on NSRDB data to find optimal parameters

In [5]:
train = cs_detection.ClearskyDetection(nsrdb.df)
train.trim_dates(None, '01-01-2015')
test = cs_detection.ClearskyDetection(nsrdb.df)
test.trim_dates('01-01-2015', None)
In [6]:
train.scale_model('GHI', 'Clearsky GHI pvlib', 'sky_status')
train.time_from_solar_noon_ratio2('Clearsky GHI pvlib')
train.scale_by_irrad('Clearsky GHI pvlib')
In [7]:
utils.calc_all_window_metrics(train.df, 3, meas_col='GHI', model_col='Clearsky GHI pvlib', overwrite=True)
In [8]:
best_params = {'max_depth': 4, 'n_estimators': 128, 'class_weight': 'balanced'}
best_params = {'max_depth': 4, 'n_estimators': 256, 'class_weight': None}
best_params = {'max_depth': 4, 'n_estimators': 256, 'class_weight': 'balanced'}
In [9]:
clf = ensemble.RandomForestClassifier(**best_params, n_jobs=-1)
clf.fit(train.df[feature_cols].values, train.df[target_cols].values.flatten())
Out[9]:
RandomForestClassifier(bootstrap=True, class_weight='balanced',
            criterion='gini', max_depth=4, max_features='auto',
            max_leaf_nodes=None, min_impurity_split=1e-07,
            min_samples_leaf=1, min_samples_split=2,
            min_weight_fraction_leaf=0.0, n_estimators=256, n_jobs=-1,
            oob_score=False, random_state=None, verbose=0,
            warm_start=False)
In [10]:
test = cs_detection.ClearskyDetection(nsrdb.df)
test.trim_dates('01-01-2015', None)
test.scale_by_irrad('Clearsky GHI pvlib')
In [11]:
%%time
pred = test.iter_predict_daily(feature_cols, 'GHI', 'Clearsky GHI pvlib', clf, 3, multiproc=True, by_day=True).astype(bool)

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.


CPU times: user 7.49 s, sys: 581 ms, total: 8.07 s
Wall time: 1min
In [12]:
vis = visualize.Visualizer()
vis.add_line_ser(test.df['GHI'], 'GHI')
vis.add_line_ser(test.df['Clearsky GHI pvlib'], 'GHI_cs')
vis.add_circle_ser(test.df[(test.df['sky_status'] == 0) & (pred)]['GHI'], 'ML clear only')
vis.add_circle_ser(test.df[(test.df['sky_status'] == 1) & (~pred)]['GHI'], 'NSRDB clear only')
vis.add_circle_ser(test.df[(test.df['sky_status'] == 1) & (pred)]['GHI'], 'ML+NSRDB clear only')
vis.show()
In [13]:
cm = metrics.confusion_matrix(test.df['sky_status'].values, pred)
vis = visualize.Visualizer()
vis.plot_confusion_matrix(cm, labels=['cloudy', 'clear'])
In [14]:
print(metrics.f1_score(test.df['sky_status'].values, pred))

0.897655453619

Train on all NSRDB data, test various freq of ground data

In [15]:
train = cs_detection.ClearskyDetection(nsrdb.df)
train.scale_model('GHI', 'Clearsky GHI pvlib', 'sky_status')
train.scale_by_irrad('Clearsky GHI pvlib')
utils.calc_all_window_metrics(train.df, 3, meas_col='GHI', model_col='Clearsky GHI pvlib', overwrite=True)
clf.fit(train.df[feature_cols].values, train.df[target_cols].values.flatten())
Out[15]:
RandomForestClassifier(bootstrap=True, class_weight='balanced',
            criterion='gini', max_depth=4, max_features='auto',
            max_leaf_nodes=None, min_impurity_split=1e-07,
            min_samples_leaf=1, min_samples_split=2,
            min_weight_fraction_leaf=0.0, n_estimators=256, n_jobs=-1,
            oob_score=False, random_state=None, verbose=0,
            warm_start=False)
In [16]:
bar = go.Bar(x=feature_cols, y=clf.feature_importances_)
iplot([bar])

1 min freq ground data

In [17]:
ground = cs_detection.ClearskyDetection.read_pickle('abq_ground_1.pkl.gz')
ground.df.index = ground.df.index.tz_convert('MST')
test = cs_detection.ClearskyDetection(ground.df)
In [18]:
test.trim_dates('10-01-2015', '10-15-2015')
In [19]:
test.time_from_solar_noon('Clearsky GHI pvlib', 'tfn')
test.scale_by_irrad('Clearsky GHI pvlib')
In [20]:
test.df = test.df[test.df.index.minute % 1 == 0]
In [21]:
print(test.df.index)

DatetimeIndex(['2015-10-01 00:00:00-07:00', '2015-10-01 00:01:00-07:00',
               '2015-10-01 00:02:00-07:00', '2015-10-01 00:03:00-07:00',
               '2015-10-01 00:04:00-07:00', '2015-10-01 00:05:00-07:00',
               '2015-10-01 00:06:00-07:00', '2015-10-01 00:07:00-07:00',
               '2015-10-01 00:08:00-07:00', '2015-10-01 00:09:00-07:00',
               ...
               '2015-10-14 23:50:00-07:00', '2015-10-14 23:51:00-07:00',
               '2015-10-14 23:52:00-07:00', '2015-10-14 23:53:00-07:00',
               '2015-10-14 23:54:00-07:00', '2015-10-14 23:55:00-07:00',
               '2015-10-14 23:56:00-07:00', '2015-10-14 23:57:00-07:00',
               '2015-10-14 23:58:00-07:00', '2015-10-14 23:59:00-07:00'],
              dtype='datetime64[ns, MST]', length=20160, freq='T')
In [22]:
pred = test.iter_predict_daily(feature_cols, 'GHI', 'Clearsky GHI pvlib', clf, 61, multiproc=True, by_day=True).astype(bool)

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.
In [23]:
for w in [3, 15, 31, 61]:
    print(w)
    pred = test.iter_predict_daily(feature_cols, 'GHI', 'Clearsky GHI pvlib', clf, w, multiproc=True, by_day=True).astype(bool)
    test.df['pred' + str(w)] = pred

3

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.


15

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.


31

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.


61

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.

/Users/benellis/duramat/clearsky_detection/utils.py:306: RuntimeWarning:

Large scaling value.  Day will not be further assessed or scaled.

/Users/benellis/duramat/clearsky_detection/utils.py:313: RuntimeWarning:

Scaling did not converge.
In [24]:
from plotly import tools as tls
In [25]:
fig = tls.make_subplots(rows=4, cols=1, shared_xaxes=True, print_grid=True)

fig['layout'].update(height=1000)

colors = {
    'blue': '#1f77b4',
    'orange': '#ff7f0e',
    'green': '#2ca02c',
    'red': '#d62728',
    'purple': '#9467bd',
    'brown': '#8c564b',
    'pink': '#e377c2',
    'gray': '#7f7f7f',
    'yellow': '#bcbd22',
    'teal': '#17becf'
}

ghi_line = {'color': colors['blue']}
ghics_line = {'color': colors['orange']}
ml_only = {'color': colors['green'], 'size': 10}
nsrdb_only = {'color': colors['red'], 'size': 10}
both = {'color': colors['purple'], 'size': 10}

for i, c in enumerate(['pred3', 'pred15', 'pred31', 'pred61']):
    legend = False
    if i == 0: legend = True
    trace0 = go.Scatter(x=test.df.index, y=test.df['GHI'], line=ghi_line, showlegend=legend, name='GHI')
    trace1 = go.Scatter(x=test.df.index, y=test.df['Clearsky GHI pvlib'], line=ghics_line, showlegend=legend, name='GHIcs')
    trace2 = go.Scatter(x=test.df[test.df[c] & ~test.df['sky_status pvlib']].index, y=test.df[test.df[c] & ~test.df['sky_status pvlib']]['GHI'], mode='markers', marker=ml_only, name='RF', showlegend=legend)
    trace3 = go.Scatter(x=test.df[~test.df[c] & test.df['sky_status pvlib']].index, y=test.df[~test.df[c] & test.df['sky_status pvlib']]['GHI'], mode='markers', marker=nsrdb_only, name='PVLib', showlegend=legend)
    trace4 = go.Scatter(x=test.df[test.df[c] & test.df['sky_status pvlib']].index, y=test.df[test.df[c] & test.df['sky_status pvlib']]['GHI'], mode='markers', marker=both, name='RF+PVLib', showlegend=legend)
    fig.append_trace(trace0, i + 1, 1)
    fig.append_trace(trace1, i + 1, 1)
    fig.append_trace(trace2, i + 1, 1)
    fig.append_trace(trace3, i + 1, 1)
    fig.append_trace(trace4, i + 1, 1)

iplot(fig)

This is the format of your plot grid:
[ (1,1) x1,y1 ]
[ (2,1) x1,y2 ]
[ (3,1) x1,y3 ]
[ (4,1) x1,y4 ]
In [ ]: